Habitat loss and fragmentation may impact animal-mediated dispersal of seed and pollen, and a key question is how the genetic attributes of plant populations respond to these changes. Theory predicts that genetic diversity may be less sensitive to such disruptions in the short term, whereas inbreeding and genetic structure may respond more strongly. However, results from studies to date vary in relation to species, context and the parameter being assessed, triggering calls for more empirical studies, especially from the tropics, where plant-animal dispersal mutualisms are both disproportionately common and at risk. We compared the genetic characteristics of adults and recruits in a long-lived palm Oenocarpus bataua in a recently fragmented landscape (<2 generations) in northwest Ecuador using a suite of 10 polymorphic microsatellite markers. We sampled individuals from six forest fragments and one nearby continuous forest. Our goal was to assess short-term consequences of fragmentation, with a focus on how well empirical data from this system follow theoretical expectations. Mostly congruent with predictions, we found stronger genetic differentiation and fine-scale spatial genetic structure among recruits in fragments compared with recruits in continuous forest, but we did not record differences in genetic diversity or inbreeding, nor did we record any differences between adults in fragments and adults in continuous forest. Our findings suggest that genetic characteristics of populations vary in their sensitivity to change in response to habitat loss and fragmentation, and that fine-scale spatial genetic structure may be a particularly useful indicator of genetic change in recently fragmented landscapes.